The development of drug therapies has been an essential approach to the treatment of infectious disease. Current treatments for AIDS rely on combinations of drugs that target essential viral enzymes, including the HIV protease. Due to error-prone replication machinery, however, populations of variant virus develop within patients, and drug treatment can select for resistant variants that render further treatment ineffective for some patients. This paradigm requires the development of a new approach to drug design. Rather than design tight-binding inhibitors to single targets, methodology is needed to design compounds that are effective against the natural target and a large set of escape mutants. This proposal aims to develop computational structure-based methodology to design "escape-resistant" inhibitors for HIV protease. The essential function of HIV protease, to cleave ten sites in gag and pol polyproteins, places a constraint on the mutational space of the enzyme (although secondary mutations in the cleavage sites have also been observed). Strategies that will be undertaken include (i) limiting the contacts that designed inhibitors make with the enzyme to those that lie within the envelope defined by the natural substrates, (ii) approaches to produce relatively small "minimal' inhibitors that contact little beyond catalytically essential residues, (iii) direct development of inhibitors to bind ensembles of known escape mutants to current therapies, (iv) development of flexible inhibitors more likely to adapt to changes in structure resulting from certain mutations, and (v) development of inhibitors to the required flap-opening-flap-closing conformational change essential for substrate entry and product exit, by targeting invariant residues in the flap-open structure. The methodology to be developed is based on "inverse design" strategies and is complementary to other computational approaches. In collaboration with other members of the proposed Program Project, inhibitor libraries will be designed, synthesized, and screened against a panel of HIV protease variants. Results at this stage alone may provide some evidence for the relative effectiveness of the different strategies above. Structures of complexes involving successful binders will be determined, and their resistance to new mutations will be assessed in selective pressure assays. The approaches being pursued here have the potential to be effective not only against HIV protease, but against a wide range of highly mutable targets.